Light emitting diodes (LEDs) are special types of semiconductor diodes first developed in the 1960s. A simplest LED consists of a p-type semiconductor and an n-type semiconductor forming a p-n junction. When an electric current passes through the junction, it creates charge-carriers (electrons and holes). In the process, an electron combines with a hole and releases energy in the form of a photon. Most of today's high efficiency LEDs have one or more layers of light emitting materials sandwiched between the p- and n-type regions to improve the light emitting efficiency. The layer structure is also used to obtain desired emission wavelengths. The basic structure of an LED device consists of a small piece of such layered material, called a die, placed on a frame or a baseboard for electrical contacts and mechanical support. The die is also encapsulated for protection.
With an LED, the wavelength of the emitted light is determined by the bandgap energy of the light emitting material. One type of material suitable for LEDs is the compound semiconductors having bandgap energies corresponding to near infrared (1R), visible or near ultraviolet (UV) light. AlGaInP (Aluminum Gallium Indium Phosphide) is one of the LED materials that exhibit high quantum efficiency (hence high brightness) and versatile colors. The bandgap of (AlxGa1-x)1-yInyP alloy system varies, depending on the x and y in the composition. The color of AlGaInP LEDs ranges from green to red. AlGaInP LEDs must be fabricated on a lattice-matching gallium arsenide (GaAs) substrate using an epitaxial growth process, such as the metalorganic chemical vapor deposition (MOCVD).
In the 1990's violet, blue and green LEDs based on gallium nitride (GaN) materials were developed. GaN is a direct bandgap semiconductor with bandgap energy of ˜3.4 eV. The electron-hole recombination in GaN leads to emission of photons at a wavelength of 360 nm, which is in the UV range. The visible wavelength LEDs (green, blue and violet) are achieved by using InzGa1-zN as the light emitting layer, sandwiched between a p-type GaN layer and an n-type GaN layer. The wavelength λ of the light emitted by the InzGa1-zN LED system varies depending on the z value. For example, for pure blue color, λ=470 nm, z=0.2. The GaN LEDs must be fabricated on a lattice-matching substrate such as sapphire or silicon carbide (SiC), again using epitaxial growth processes such as MOCVD.
Great efforts have been made to produce white LEDs as a replacement for conventional lighting sources. Currently, white color LEDs can be accomplished in various ways:                (1) Putting discrete red, green and blue LEDs in a “lamp” and use various optical components to mix light in red, green and blue colors emitted by those discrete LEDs. However, because of the different operating voltages for LEDs of different colors, multiple control circuits are required. Furthermore, the lifetime of the LEDs is different from one color to another. Over time the combined color would change noticeably if one of the LEDs fails or is degrading.        (2) Partially converting light in short wavelengths to light in the longer wavelengths using phosphors. One of the most common ways is to cover a yellowish phosphor powder around a blue InGaN LED chip. The phosphor powder is usually made of cerium doped yttrium aluminum garnet (YAG:Ce) crystals. Part of the blue light emitted by the InGaN LED chip is converted to yellow by the YAG:Ce. However, the “white” light so produced contains mainly two colors: blue and yellow. Such a light source is usually used as indicator lamps.        (3) Using UV light produced by very short-wavelength LEDs to excite phosphors of different colors in order to produce light in three basic colors. The drawback of this method is that the lifetime of the UV LEDs is relatively short. Furthermore, UV radiation from the LEDs can be a health hazard, as most of commonly used encapsulation materials today are not effective in blocking UV radiation.        
There have been numerous attempts in developing white LED light sources with higher efficiency and better chromaticity. Guo et al (“Photon-Recycling for High Brightness LEDs”, Compound Semiconductor 6(4) May/June 2000) suggests the concept of photon recycling in producing high brightness white LEDs. Photon recycling is a process by which short wavelength photons are absorbed by an emitter material, which re-emits photons of long-wavelengths. In principle, photon recycling semiconductor (PRS) LEDs can efficiently produce white light up to 330 lumen/watt. However, the drawback of PRS-LEDs is their extremely low color-rendering index.
The dual-color PRS-LED, as disclosed in Guo et al., consists of a primary light source and a secondary light source. The secondary light source has a secondary light-emitting layer. The primary light source is used to produce blue light. The produced blue light is directed to the secondary emitting layer so that part of the blue light is absorbed in order to produce yellow light in the re-emitting process. In principle, the dual-color photon production in PRS-LEDs is analogous to the phosphor coated LED. However, unlike the phosphor coated LED, the secondary light source consists of a fluorescent semiconductor material, AlGaInP, directly bonded to the primary light source wafer. It is therefore possible to produce dual-color LED chips on a wafer. FIG. 1 shows the structure of a PRS-LED, according to Guo et al. As shown in FIG. 1, the PRS-LED 1 consists of a transparent substrate 18 made of sapphire. The primary light source and the secondary light source are disposed on opposite sides of the sapphire substrate. The primary light source comprises a p-type GaN layer 12, an active layer 14 made from InGaN and an n-type GaN layer 16. These layers are epitaxially grown on the sapphire substrate 18. The secondary light source LED consists of mainly a layer of AlGaInP 22. The AlGaInP layer is epitaxially grown on a GaAs substrate (not shown) and then bonded to the sapphire substrate 18 using a bonding material 20. The GaAs substrate is subsequently removed by chemically-assisted polishing and selective wet etching. After the primary light source layers are patterned, an n-type contact 36 made of Al is deposited on a section of the n-type GaN layer 16, and a p-type contact 32 of Ni is deposited on a section of the p-type GaN layer 12.
The primary light output is produced in the active region 14 by current injection, and the wavelength of the primary light is approximately 470 nm. In operation, a portion of the photons emitted by the primary light source is absorbed by the AlGaInP layer 22 and then re-emitted (or recycled) as photons of a longer wavelength. The composition of the AlGaInP layer 22 can be selected such that the re-emitted light is at the wavelength of 570 nm (yellow). Because the colors of the light produced by the primary light source and the secondary light source are complementary, the combined light output appears white to the human eye. In such a PRS-LED structure, while the white light contains emission peaks of 470 nm (blue) and 570 nm (yellow), no red light (˜650 nm) is emitted.
The mixed light produced by the aforementioned methods may appear white to the human eye. However, the mixed light does not have the required chromaticity as required in a quality color display, such as an LCD display.
Thus, it is advantageous and desirable to provide a method to produce a semiconductor light source containing wavelength components in RGB.